EsportsBot AI

Inspiration

We noticed that navigating esports data often requires digging through multiple websites, piecing together player stats, tournament earnings, and game rankings. Our inspiration was to create a streamlined, AI-powered chatbot that instantly provides high-level insights without manual research.

What it does

EsportsBot AI uses a Large Language Model (LLM) to interpret user questions, then dynamically fetches real-time esports earnings data from the official Esports Earnings API. It summarizes the findings into clear, conversational responses—covering player rankings, tournament results, and top earners for specific games.

How we built it

LLM Integration: We used Google’s Gemini 2.0 Flash to parse user queries and determine the appropriate API calls (e.g., LookupPlayerById, LookupHighestEarningPlayersByGame). Esports Data API: We integrated with the Esports Earnings API to fetch live data about players, teams, games, and tournaments. Data Summarization: The LLM also synthesizes fetched information into user-friendly explanations. UI/UX Layer: A Streamlit web app for a responsive, interactive user experience. Challenges we ran into

API Constraints: The Esports Earnings API sometimes has limited data or SSL/certificate issues, which required careful error handling and workarounds. LLM Prompt Engineering: Crafting prompts that accurately guide the LLM to generate correct JSON for API calls and produce concise summaries. Game ID Lookups: Managing game IDs dynamically without hardcoding, including scraping or CSV-based solutions.

Accomplishments that we're proud of

End-to-End Integration: Seamlessly bridging LLM capabilities with real-time esports data to deliver context-aware answers. User-Centric Design: A simple chatbot interface that speaks in plain language, hiding the complexity of multiple API endpoints behind the scenes. Robust Architecture: Error handling, fallback logic, and special cases (e.g., missing data) to ensure consistent, reliable responses. What we learned

LLM-Oriented Development: Best practices around prompting, chaining, and structuring responses for real-world use cases. Practical Data Handling: Balancing the power of semantic queries with concrete data constraints from external APIs. Scalability & Caching: Strategies to handle repeated or complex queries without exceeding API rate limits. What's next for EsportsBot AI

Advanced Analytics: Introducing player performance trends, deeper tournament breakdowns, and real-time stats visualizations. User Personalization: Providing custom alerts or recommendations based on user-tracked players and teams.

Extended LLM Features: Building memory into the conversation flow, enabling follow-up questions like “Who else is in the top 10 for that tournament?” without restating context. Global Deployment: Scaling to handle large numbers of concurrent users and potentially integrating additional esports data sources.